oh-my-opencode
Fail
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill contains several instructions that download content from third-party domains and pipe it directly to a shell for execution. Examples include 'curl -fsSL https://opencode.ai/install | bash' in the SKILL.md and installation files, and 'bash -c "$(curl -fsSL https://raw.githubusercontent.com/code-yeongyu/oh-my-opencode/refs/heads/master/docs/guide/installation.sh)"' in the installation guide.
- [COMMAND_EXECUTION]: The skill promotes the use of privilege escalation to perform system-level operations such as installing global packages. Evidence found in oh-my-opencode-installation/SKILL.md includes instructions to run 'sudo pacman -S opencode' and 'sudo npm install -g oh-my-opencode'.
- [EXTERNAL_DOWNLOADS]: The skill requires downloading and installing Node.js packages from external registries. Detected packages include 'oh-my-opencode', '@opencode-ai/plugin', and 'opencode-ai'.
- [PROMPT_INJECTION]: The skill reads and processes untrusted data from project-level and user-level configuration files without providing safety boundaries or sanitization, creating a surface for indirect prompt injection. Ingestion points: .opencode/oh-my-opencode.jsonc and ~/.config/opencode/oh-my-opencode.jsonc. Capability inventory: Full shell command execution via the 'bash' and 'opencode' tools. Boundary markers: None identified. Sanitization: None identified.
Recommendations
- HIGH: Downloads and executes remote code from: https://opencode.ai/install - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata